Personal characteristics of vocational rehabilitation applicants: Findings from the Survey of Disability and Employment

2.1 Questionnaire design

Our selection of survey content was motivated and guided by the social model of disability (Nagi, 1965; Verbrugge & Jette, 1994) and the International Classification of Functioning, Disability and Health model (World Health Organization, 2001). Both frameworks suggest that many domains interact to determine disability and employment outcomes and that disability itself is a dynamic function of health, functioning, and the physical and social environment. We designed the SDE to capture data in the multiple domains that these models predict might affect employment outcomes. Whenever possible, we adapted survey items from existing surveys. When we could not identify existing questions for the survey, we wrote new questions. Table 1 provides an overview of the instrument domains and sources.

The domains discussed in this paper include (1) personal characteristics, such as demographics and education; (2) health conditions and the physical and mental characteristics that underlie disability; (3) employment and social participation, such as an individual’s family, community and employment relationships, and employment experiences; and (4) environmental factors including accessibility, transportation, the local economy, public policies, and geography. Barriers and facilitators to employment can appear in any of these different domains, and they may present themselves through interactions across domains as well.

The SDE includes a comprehensive set of health and disability questions designed to provide a more comprehensive understanding of these domains than is available in most existing surveys. Disability measurements in most existing surveys do not yield in-depth information on the health issues and functional limitations experienced by individuals with disabilities (Livermore et al., 2011). For example, the six-question sequence first included in the Current Population Survey (CPS) and now required in all national surveys has been criticized for missing many individuals with disabilities (Burkhauser et al., 2014). Measures related to psychiatric and cognitive impairments in existing surveys are particularly poor, in part because questions have not been developed but also because these groups have been underrepresented in national surveys due to issues of consent and mode of survey (Livermore et al., 2011). The SDE first asks about health with two open-ended questions about the respondent’s health issues or limitations and the causes of these issues. To facilitate comparison with data from national surveys, we followed with a series of “yes/no” questions on particular functional impairments based on the CPS six-question sequence and a self-rated health question. We also included a question on pain, which is not assessed by national surveys despite the fact that pain may be playing an increasing role in the growth of disability programs (Autor & Duggan, 2006).

We also constructed the employment section to be more comprehensive than most existing surveys used to study employment of individuals with disabilities. Surveys such as the CPS and the American Community Survey that are regularly used to estimate employment rates collect employment-related information only from those who are currently working. Although the SDE is cross-sectional like those surveys, respondents are asked about characteristics of both current and prior jobs, such as hours of work, occupation, and industry. Additionally, they are asked questions about their employment history, including workplace accommodations and the timing and reasons for job separation.

Stakeholders, including state VR agency staff, provided feedback on the draft instrument. The project’s scientific review committee also reviewed the draft survey instrument. We collated all feedback and presented it to the investigator team for review. After the team agreed, we incorporated the feedback into the survey. The institutional review board at the Kessler Foundation reviewed and approved the protocols for human subjects.

2.2 Study population

The target population for the survey was working age individuals, ages 25 to 60 years, with disabilities who were interested in employment and who had recently applied for state or federal VR services through a state VR agency. We sampled VR applicants for a number of reasons. First, because the VR program is limited to individuals with significant disabilities, sampling from this population avoids costly prescreening for disability from a larger population. Second, because VR is designed to help with employment, VR applicants have actively demonstrated that they are employment focused. Third, VR applicants are likely to have more recent employment experiences than other populations with disabilities, such as Social Security beneficiaries, because eligibility for VR services is not contingent on being out of the labor force.

With endorsement from the Rehabilitation Services Administration (RSA), the study funder, and the Council of State Administrators of Vocational Rehabilitation, we recruited three agencies to participate in the study: the New Jersey Division of Vocational Rehabilitation Services (henceforth “New Jersey”), the Mississippi Department of Rehabilitation Services (“Mississippi”), and Opportunities for Ohioans with Disabilities (“Ohio”). Mississippi and Ohio are “combined” agencies that serve both the blind and visually impaired and other individuals with disabilities, whereas New Jersey is a “general” agency that does not serve the blind and visually impaired. After implementing data protections and agreements with Mathematica, the agencies sent Mathematica files with records on individuals who applied for VR services between August 15 and December 15, 2014.

2.3 Survey response

A total of 2,804 applicants gave consent and participated in the survey. The survey had an overall response rate of 56.6%, which we calculated using the American Association for Public Opinion Research’s Response Rate 3 (American Association for Public Opinion Research, 2015). The respondents were almost equally distributed across Mississippi, New Jersey, and Ohio, with a slightly smaller share coming from Mississippi. Table 2 provides additional information on survey outcomes by agency. Interviews were conducted directly with the sampled respondent 96% of the time; 100 interviews (4%) were conducted using a proxy respondent. In addition, roughly 1% of interviews (n = 24) were conducted using assistive technologies. We calculated sampling weights as the inverse of the probability of selection, with adjustments for differential nonresponse. In addition, we performed a nonresponse bias analysis to evaluate how well our nonresponse adjustments accounted for the differences between respondents and the full set of applicants in the sample frame. We found that response rates were higher among older applicants in all three states, among Blacks in Ohio and Mississippi, and among non-Hispanics in New Jersey. Response rates were lower among those with psychosocial or behavioral disabilities in Mississippi. Nonresponse adjustments to the weights correct for most of these differences.

3.1 Generalizability and sample characteristics

To assess the extent to which SDE findings could be generalized to the VR population, we made comparisons with the national population of VR clients using the 2013 RSA-911 case closure file (Table 3). Although some differences exist between the two, we find the SDE respondents largely approximate the VR population; thus findings from the SDE should be generalizable to the national VR population, with some exceptions. The two groups are very similar by age and educational attainment; however, a slightly higher weighted share of SDE respondents are female or non-White, and a slightly lower share report psychiatric disabilities. The one striking difference between the SDE sample and the VR population is that whereas 17% of the VR population was employed at application, roughly a third of SDE respondents report being employed. These differences likely reflect a mix of factors, including differing SDE response rates across demographic and impairment groups and economic differences over time and across states. Employed VR applicants may also have been more willing to participate in the survey, given the survey focus on employment.

3.2 Health and disability

Table 4 contains information about health and disability. As we mentioned, health and disability are multidimensional, and no single question fully captures their complexity. As a result, we present findings across SDE questions that ask about health and disability in different ways.

Respondents were prompted with an open-ended question to describe the health conditions or limitations that affect their daily activities. We coded their responses into multiple categories as appropriate. The most commonly reported impairments included physical impairments (46%), followed by psychiatric conditions (32%), mobility impairments (30%), and sensory impairments (19%). Despite the fact that VR serves only individuals with disabilities, almost 5% of applicants reported that they did not have a condition that affected their daily activities.

Responses to questions about functional limitations highlight the fact that those seeking VR services face a range of limitations that may not align perfectly with those expected based on their reported impairments. As described above, the SDE includes a seven-question functional limitation sequence based on the six disability questions from the CPS and an additional question on chronic pain. Twenty-three percent of our respondents reported difficulty seeing, even with eyeglasses, a figure that is substantially higher than the 8% who mentioned blindness or vision impairments when asked to describe their health condition. Similarly, 54% reported difficulty concentrating, remembering, or making decisions, a substantially higher rate than, but not perfectly comparable to, the 39% who mentioned intellectual disability, memory impairments, and/or psychiatric conditions. Half the sample stated that they regularly experience pain.

The SDE also asks a series of questions to assess executive functioning. Executive functioning, broadly described as a set of mental skills that help one get things done, has been found to be predictive of positive outcomes among rehabilitation patients (Hanks et al., 1999). Overall, SDE respondents’ self-perceptions suggest high levels of executive functioning. Eighty-six percent strongly agree that they are or would be a reliable worker; 78% strongly agree that they persevere until a task is done; while 20% strongly agree that they are easily distracted.

Although disability and health status can be closely related, for many of our respondents they are not. Poor health may independently influence employment opportunities. As a result, SDE respondents separately rated their health excellent (8%), very good (18%), good (33%), fair (30%), or poor (11%). The range of self-rated health status again displays the diversity of VR applicants.

A minority of those seeking VR services have experienced a recent onset of disability (10% within 1 year (Fig. 1) and an additional 19% within 2 to 5 years). Those with recent disability onset may present with VR needs that are different from the needs of those whose disability onset was far in the past (5% of our sample reported they had had their disability since birth, and an additional 27% had had their disability for more than 20 years). How long an individual has had a disability or the age at which he or she first experienced the disability can influence the person’s needs in various ways. For example, a new disability may be disruptive but over time the person may be able to adapt to it. Alternatively, over time the severity of a person’s disability may worsen, or the person’s resources for adapting to it may be exhausted (Jenkins & Rigg, 2004).

3.3 Employment and social participation

Employment has been a part of most applicants’ lives. Three-quarters of applicants worked prior to disability onset, three-quarters have worked since onset, and over a third are employed at application for VR services (Table 5). Most respondents work or have worked a nontrivial number of hours; respondents reported working a median of 32 hours per week and 25% reported working 40 hours or more per week. However, we found substantial heterogeneity in respondents’ employment history. A majority (more than 65%) have worked at some point in the past 2 years, but 25% have not worked for 5 or more years, and 2.5% have never worked.

Respondents to the SDE have worked in a wide range of industries, with the most common being in the service sector: health care, retail trade, administrative support, and food services. Together, these jobs account for 57% of the jobs held currently or most recently held. The next most common industry is manufacturing, with 9.5% of jobs.

Consistent with substantial past employment for many respondents, a vast majority of SDE respondents stated that it is very important or extremely important that they work. Figure 2 shows how this response varies by years since the respondent was employed. Even among those who have not worked in a decade, 70% say work is very important or extremely important. That share rises to 93% among those currently employed.

In addition, SDE respondents report varying levels of social participation. Researchers have theorized that social participation serves as a facilitator to employment, and research on the general population of persons with disabilities finds that higher levels of social capital are associated with positive employment outcomes (Araten-Bergman & Stein, 2014; Brucker, 2015). Table 5 shows that 42% of VR applicants say they engage in social activities at least once a week; 27% do so less frequently but at least monthly; and 18% say they never engage in social activities.

3.4 Barriers and facilitators to participation

Although work is important to the vast majority of VR applicants, the fact that they have applied for VR services suggests that they are facing challenges to employment. A primary challenge to work is poor health. When asked why they are not currently working, two-thirds said that their health prevents it (Table 6). When asked about why or how their health prevents work, 80% said their condition becomes unpredictably better or worse (Fig. 3). Almost 60% mentioned pain, lack of physical energy, and the time required to take care of health. Roughly 40% each mentioned that work was too stressful or that work was harmful to their health. Lastly, 10% mentioned that their health led to disagreements with coworkers.

In addition to health, respondents cited many other supply-side and personal reasons for not currently working (Table 6). More than half the sample said they could not find a job and almost half said they felt employers would not give them a chance and they were discouraged by previous attempts at working. Almost a quarter said that their family and friends discouraged them from working. Other barriers mentioned included lacking the skills for a job (43%) and transportation and workplace accessibility (30% each). Almost a quarter mentioned fear of losing their cash benefits or Medicaid.

Respondents who were not employed at the time of the survey also reported why they left their most recent job. Roughly a third of the respondents mentioned their health or being fired as the reason they left their job (Table 6). While respondents mentioned many other reasons, including childcare or transportation issues, or that they did not like the job, less than 10% of the sample mentioned any of these other reasons.

Despite the many employment barriers respondents faced, they frequently reported certain factors as facilitators to employment. Both respondents currently working and those formerly working reported these facilitators. Some of these accommodations, such as transportation (21%) and modifications to space and equipment (17%), pertain to physical barriers. Other accommodations rely on assistance from coworkers (33%), job coaches (42%), or personal care attendants (10%). A third category of accommodations involves alterations of job duties or hours, with the most prevalent accommodations being a flexible schedule (reported by 49%) and modified work duties (reported by 28%).